Pediatric emergency department organization and social care practices among U.S. fellowship programs

Abstract

Background

Social care has become increasingly relevant to the emergency physician and includes activities that address health‐related social risk and social needs. The literature has consistently documented substantial health care provider challenges in incorporating social care into routine practice. Yet, interventions on the health care organizational level hold promise to bring about more widespread, sustainable impact.

Methods

This study was a subanalysis of the 2021 National Social Care Practices Survey data set among pediatric emergency medicine (PEM) program directors (PDs) and fellows. The purpose was to investigate the association between health care organizational factors and PEM physician social care practices and perceptions among PEM PDs and fellows. We performed binary and ordinal logistic regressions of organizational factors and five specific PEM physician social care perspective and practice outcomes.

Results

The sample population included 153 physicians—44 PDs (49% response rate) and 109 fellows (28%). PDs and fellows with access to a social care systematic workflow in their pediatric emergency department (PED) had higher odds of comfort assessing social risk (odds ratio [OR] 2.1%, 95% confidence interval [CI] 1.1–4.0), valuation of social care (OR 3.2, 95% CI 1.3–7.9), preparedness to assist families (OR 2.4, 95% CI 1.1–5.2), screening tendency (OR 2.2, 95% CI 1.1–4.5), and ability to refer to community resources (OR 2.3, 95% CI 1.2–4.6). A similarly directed, but less pronounced pattern was noted with access to a community resource database for referrals and 24‐h access to a social worker in the PED.

Conclusions

PED organizational factors—particularly access to a social care systematic workflow—appear positively associated with PEM physician practices and perceptions of social care delivery. Further research is under way to advance understanding of PEM training factors in social care.

INTRODUCTION

Despite a surge of literature demonstrating significant unaddressed social needs among families in the emergency department (ED) and their impact on child health, there remains no standard organizational practice of social care in the ED. ¹ , ² , ³ , ⁴ Social care includes activities that address health‐related social risk and social needs, whereas social risk encompasses social conditions associated with poor health, social needs are the specific conditions in which patients and their families request assistance. ⁵ We demonstrated in a preceding study that the majority of a national cohort of pediatric emergency medicine (PEM) fellows and program directors (PDs) highly valued social care yet lacked organizational structure to address common social risk such as housing instability, food insecurity, access to public benefits, and transportation assistance. ⁶

In 1999, Dr. James Gordon eloquently theorized the ED as a social welfare institution and societal resource functioning as a “social triage” and an extension of existing community resources. ⁷ While screening for social risk has become a norm among pediatricians in the office setting, common practice in the emergency setting falls well short of a coordinated, systematic approach integrating social care into clinical care. ⁵ , ⁸ , ⁹ , ¹⁰ In the ED, effectively addressing social needs poses systemic dilemmas in workforce training, resource allocation, clinical operations, and workflow. ² Successful models have been published by individual medical centers; however, the majority of pediatric emergency departments (PEDs) lack comprehensive structure to support two primary social care functions: awareness and assistance. ³ , ⁵ , ⁶ , ¹¹ , ¹² Awareness includes health care system activities that identify social risk, while assistance refers to those that reduce social risk by connecting patients with relevant community resources. ⁵

Even less is known regarding PEM physician perspectives of social care structure and operations. Various research has documented physicians' strong support of social care integration into emergency medical care, with limited time and training as principal barriers at the physician level. ⁶ , ¹³ , ¹⁴ , ¹⁵ Lessons learned from pediatric and adult emergency medicine social care networks suggest that effective and sustainable strategies recognize these physician‐level limitations in pursuit of organizational solutions. ³ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ In the ED, these macrolevel interventions, or organizational factors, aim to operationalize social risk awareness and assistance by integrated workflow approaches, community referral databases, and leveraging social work services. ⁵ , ¹⁹ In this study, our purpose was to assess the association between health care organizational factors and PEM physician social care perception and practices among PEM PDs and fellows in the United States.

METHODS

This cross‐sectional study utilized data collected from the 2021 National Social Care Practices Survey. ⁶ PEM PDs and fellows in all accredited PEM fellowship training programs in the United States were included in recruitment for the national survey, which assessed social care knowledge, perspectives, and training. ⁶ The study was exempt for review by the primary author's institutional review board.

Variable definitions

Since a best‐practice approach to PED‐based social care has not yet been established, PED organizational factors in this study were broadly defined to encompass a range of practices and models across PEDs nationally. Factors of interest included physician access to (1) a social care systematic workflow, (2) a community resource database, and (3) 24‐h access to a social worker. We defined social care systematic workflow as an organized approach to consistently provide social risk awareness to the majority of patients seen in the ED. Access to a community resource database was understood as a keystone of social needs assistance activity. ⁶ Twenty‐four‐hour access to a social worker included both in‐person and telephone consults with a social worker managing the ED.

Outcomes of interest among the sample of PEM PD and fellow physicians included comfort assessing patient/families for social needs, perceived value of social care, preparedness to provide social care, screening tendency, and social needs referral activity. Comfort, value, and preparedness were assessed in a 5‐item Likert scale including a neutral option which were categorized as higher‐scale, neutral, and lower‐scale options (i.e., favorable, neutral and unfavorable responses). Responses for social needs screening tendency were dichotomized to “do screen” (screened for social needs 1%–100% of the time) and “do not screen.” In regard to referral activity, participants were asked about their ability to independently (i.e., without the aid of a social worker or community navigator) refer patients/families to community resources for housing instability, food insecurity, accessing public benefits, assistance paying utilities, transportation assistance, or immigration assistance with a yes/no response. Here, participants were grouped into “do refer” if they reported an ability to independently refer patients/families to one or more areas of social need. Otherwise, participants were grouped into “do not refer.”

Statistical analysis

We first computed frequency distributions of demographic variables and organizational variables (systematic workflow, community resource database, and 24‐h access to social worker) overall and by each specified outcome (comfort, value, preparedness, screening, and referral activity). We then performed unadjusted ordinal logistic regressions to assess the relationship between organizational factors and physicians' comfort, value, and preparedness with social care. These models allow for evaluation of a dependent variable with greater than two levels in a natural order (i.e., very prepared, prepared, neutral, unprepared, very unprepared). For each dependent variable, we combined the higher two levels (i.e., very prepared and prepared) and the lower two levels (i.e., unprepared and very unprepared) to facilitate comparison between each other and the neutral value. We assessed the proportional odds assumption, which assumes there is a natural order in the outcome using an alpha of 0.05. Finally, we conducted unadjusted binary logistic regression models for organizational factors on screening tendency and social needs referral activity. All models were conducted separately for each organizational factor and outcome. Missing data for the sample was minimal overall, and it was assumed that any missing data was missing completely at random. All analyses were conducted using SAS software Version 9.4 of the SAS System for Windows (SAS Institute Inc.).

RESULTS

Of the 89 PEM PDs contacted, 44 (49%) participated and 109/383 (28%) PEM fellows participated. Demographic characteristics of participants as well as a distribution of ED structural factors and physician social care practices are reported in a preceding descriptive study utilizing the National Social Care Practices Survey data set. ⁶ Participating PDs and fellows were geographically diverse, from academic institutions in the West/Northwest (21%), Midwest/Central (29%), South/Southeast (24%), and East/Northeast (26%) United States.

Comfort with asking about social risk

Among fellows and PDs who reported access to a social care systematic workflow, 55% reported feeling comfortable asking patients and families about social risk, 29% neutral, and 16% uncomfortable (Table 1). In comparison, 42% of those without access to a workflow reported feeling comfortable, 20% neutral, and 38% uncomfortable. Those who reported access to a social care systematic workflow were 2.1 times more likely to have higher levels of comfort compared to those without a workflow (95% confidence interval [CI] 1.1–4.0). Fifty‐four percent of those with 24‐h access to a social worker reported comfort, 23% neutral, and 23% uncomfortable with asking patients/families about social risk. The odds of higher levels of comfort with social risk assessments for those with 24‐h access to a social worker were 3.6 times that of those without 24‐h access (95% CI 1.8–7.3).

TABLE 1.

Organizational factors and social risk assessment comfort, PEM fellows and PDs, combined (2021, N = 153)

	Comfort asking patients/families about social risk			Unadjusted ordinal logistic regression
	Comfortable, n (%)	Neutral, n (%)	Uncomfortable, n (%)	OR (95% CI)	Proportional odds, p‐value
Organizational factors
Social care systematic workflow
Yes	28 (55)	15 (29)	8 (16)	2.1 (1.1–4.0)	0.07
No/do not know (ref)	43 (42)	20 (20)	39 (38)	—
Community resource database
Yes	31 (53)	18 (30)	10 (17)	1.1 (0.5–2.2)	0.2
No/do not know (ref)	40 (43)	17 (18)	37 (39)	—
24‐h access to social worker
Yes	61 (54)	26 (23)	26 (23)	3.6 (1.8–7.3)	0.9
No/do not know (ref)	10 (25)	9 (22)	21 (53)	—

Abbreviations: PD, program director; PEM, pediatric emergency medicine; ref, reference.

Assigned value of social care

The vast majority (86%) of respondents with access to a social care systematic workflow reported a high value of providing social care to patients/families in the ED, while 8% reported feeling neutral, and 6% reported low value (Table 2). The odds of assigning higher value to social care among those with access to a systematic workflow were 3.2 times compared to those who did not have a systematic workflow (95% CI 1.3–7.9).

TABLE 2.

Organizational factors and social care value, PEM fellows and PDs, combined (2021, N = 153)

	Assigned value of social care			Unadjusted ordinal logistic regression
	Important, n (%)	Neutral, n (%)	Not important, n (%)	OR (95% CI)	Proportional odds, p‐value
Organizational factors
Social care systematic workflow
Yes	44 (86)	4 (8)	3 (6)	3.2 (1.3–7.9)	0.6
No/do not know (ref)	68 (67)	13 (13)	21 (20)	—
Community resource database
Yes	27 (85)	3 (9)	2 (6)	2.4 (0.8–6.7)	0.5
No/do not know (ref)	85 (70)	14 (12)	22 (18)	—
24‐h access to social worker
Yes	86 (76)	13 (12)	14 (12)	1.8 (0.9–4.0)	0.3
No/do not know (ref)	26 (65)	4 (10)	10 (25)	—

Abbreviations: PD, program director; PEM, pediatric emergency medicine; ref, reference.

Preparedness to assist with social needs

Overall, only a minority of fellows (7%) and PDs (9%) reported feeling prepared to assist families with social needs. ⁶ A higher proportion of fellows and PDs with access to a social care systematic workflow (14%) reported feeling prepared to personally assist (i.e., without the help of a social worker/navigator) families whose social needs arise during their ED visit compared to those without a systematic workflow (5%; Table 3). The odds of higher levels of preparedness among those with access to a systematic workflow were 2.4 times that of those without access to a systematic workflow (95% CI 1.1–5.2). Of those with access to a community resource database, 13% reported feeling prepared, 25% neutral, and 62% unprepared. Of those without access to a community resource database, 7% reported feeling prepared, 12% neutral, and 81% unprepared. The odds of higher levels of preparedness for those with access to a community resource database were 2.5 times compared to those without access (95% CI 1.1–5.7).

TABLE 3.

Organizational factors and social care preparedness, PEM fellows and PDs, combined (2021, N = 153)

	Preparedness to assist with social needs			Unadjusted ordinal logistic regression
	Prepared, n (%)	Neutral, n (%)	Not prepared, n (%)	OR (95% CI)	Proportional odds, p‐value
Organizational factors
Social care systematic workflow
Yes	7 (14)	10 (20)	34 (66)	2.4 (1.1–5.2)	0.6
No/do not know (ref)	5 (5)	13 (13)	84 (82)	—
Community resource database
Yes	4 (13)	8 (25)	20 (62)	2.5 (1.1–5.7)	0.7
No/do not know (ref)	8 (7)	15 (12)	98 (81)	—
24‐h access to social worker
Yes	10 (9)	16 (14)	87 (77)	1.1 (0.5–2.6)	0.4
No/do not know (ref)	2 (5)	7 (17)	31 (78)	—

Abbreviations: PD, program director; PEM, pediatric emergency medicine; ref, reference.

Screening tendency

Among fellows and PDs with access to a social care systematic workflow, 73% reported personally performing social needs screening over their last five shifts compared to 55% of those without such a workflow, crude odds ratio (cOR) 2.2 (95% CI 1.1–4.5; Table 4). There appeared to be less of a difference among those with and without access to a community resource database or 24‐h access to a social worker.

TABLE 4.

Organizational factors and screening tendency, PEM fellows and PDs, combined (2021, N = 153)

	Screening tendency		Unadjusted logistic regression
	Do screen, n (%)	Do not screen, n (%)	OR (95% CI)
Organizational factors
Social care systematic workflow
Yes	37 (73)	14 (27)	2.2 (1.1–4.5)
No/do not know (ref)	56 (55)	46 (45)	—
Community resource database
Yes	19 (59)	13 (41)	0.9 (0.4–2.1)
No/do not know (ref)	74 (61)	47 (39)	—
24‐h access to social worker
Yes	69 (61)	44 (39)	1.1 (0.5–2.2)
No/do not know (ref)	24 (60)	16 (40)	—

Abbreviations: PD, program director; PEM, pediatric emergency medicine; ref, reference.

Referral ability

Survey participants were asked about their ability to assist families with community resource referrals in any one of six specific social needs if they did not have immediate access to a social worker or community navigator (Table 5). These included housing instability, food insecurity, accessing public benefits, assistance paying utilities, transportation options, and immigration assistance. A higher proportion of fellows and PDs with access to systematic workflow for screening (61%) reported the ability to perform community referrals compared to 39% of those without a systematic workflow (cOR 2.3, 95% CI 1.2–4.6). Again, there appeared to be less of a difference among those with and without access to a community resource database or 24‐h access to a social worker.

TABLE 5.

Organizational factors and referral activity, PEM fellows and PDs, combined (2021, N = 153)

	Referral activity		Unadjusted logistic regression
	Do refer, n (%)	Do not refer, n (%)	OR (95% CI)
Organizational factors
Social care systematic workflow
Yes	31 (61)	20 (39)	2.3 (1.2–4.6)
No/do not know (ref)	41 (40)	61 (60)	—
Community resource database
Yes	19 (59)	13 (41)	1.9 (0.9–4.1)
No/do not know (ref)	53 (44)	68 (56)	—
24‐h access to social worker
Yes	51 (45)	62 (55)	0.7 (0.4–1.5)
No/do not know (ref)	21 (52)	19 (48)	—

Abbreviations: PD, program director; PEM, pediatric emergency medicine; ref, reference.

DISCUSSION

In this study we identified a generally positive‐directed relationship between organizational factors and physician perspective of social care as well as awareness and assistance practices. Specifically, access to a social care systematic workflow, community resource database, and ED social worker appeared to have some positive association, with a systematic workflow most consistently related to physician perspective and practice of social care. This research suggests that organizational support may leverage the PEM physician's ability to provide social care among patients and families seen in the PED. However, regardless of the impact of organizational factors, the vast majority of our study population (80% of fellows and 70% of PDs) reported feeling unprepared to assist families with social needs, which draws the role of the physician into question.

It has been argued that the individual ED physician may have the greatest impact not by screening each patient for social risk but rather by sharing a responsibility to leverage the patient encounter for “maximum total health benefit,” that is, by improving health outcomes through simultaneous medical and social care. ⁷ The fact that in our research there is a discordance between PEM physicians' overwhelmingly positive outlook of social care yet poor preparedness in practice brings attention to upstream ED structural features that support social care delivery. These features include, but are not limited to, a clear organizational mission to address the social needs and risk factors of patients/families (value), workforce training including screening practices that avoid patient/family stigmatization (comfort with asking about social risk), and a network of community resources (screening and referral ability and capacity). These findings were concordant between PEM faculty and fellow opinion, which points away from generational or training differences among these groups and toward ED infrastructural factors in workforce organization and training. Indeed, the ED has been long understood as a de facto social safety net, yet its strategic coordinated development as an effective societal resource for identification of basic social risk and extension of existing community resources has gone widely unrealized. ² , ⁷ , ²¹ , ²² , ²³ Most recently, targeted social intervention in the ED has expanded and process improvement models have been developed for social risk assessment, referral, and service coordination but are yet to be broadly adapted into practice. ² , ³ , ⁷ , ¹² , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

This national research identifies a connection between physician social risk awareness and assistance practices (screening and referral activity, respectively) with ED organizational capacity. PEM PDs and fellows were more likely to perform social risk screening and referral when there was a social care systematic workflow in place. In the absence of such organized support, physician screening for social risk is expectedly low and inconsistently implemented. ⁴ , ¹⁵ , ¹⁶ , ²⁴ , ²⁵ , ²⁶ Limited time and training are the primary individual provider‐level factors that have been shown to consistently hinder screening and referral practices. ¹³ , ¹⁴ , ¹⁵ , ²⁷ , ²⁸ , ²⁹ Best organizational practices in social risk screening have been recently developed but not widely adapted to the emergency setting. ¹⁶ , ¹⁸ In practice, infrastructure building should precede social risk screening effort and address a series of early key steps, including clearly defining key social factors driving poor health outcomes in a local population (patient needs assessment), an evaluation of existing community resources (asset mapping), and an evaluation of capacity and readiness to enable measurable and sustainable change (workforce needs assessment). ³

Physician perspectives of social care (comfort with assessing social risk, assigned value to social care, and preparedness) were found to be positively associated with ED organization, especially with having a systematic workflow in place. Again, responses were concordant between PEM fellows and PDs, highlighting the relevance of institutional and administrative strategies that make social care a part of routine medical care rather than training or generational differences between the two groups. Conversely, low organizational reinforcement of health care worker capacity to identify and provide care for patients with social need has been associated with a high burnout rate among physicians and nurses alike. ³⁰ The positive association between physician perspective and ED organization among this study's population also likely reflects integration of social care at multiple levels, including into a value‐based ED mission statement, logistically into ED throughput, and operationally into ED workforce training and community partner networking. While multiple studies have documented patient and family preference for electronic, self‐administered screening, ²⁰ , ³¹ , ³² providers including physicians, nurses, social workers, case managers, and administrators share favorable views of systematic universal screening and referral platforms incorporated into clinical workflow. ⁶ , ¹³ , ¹⁶ , ³³ , ³⁴ As ED social care process models continue to be developed nationally and are assessed for feasibility and sustainability, we argue that physician perspective, and likely that of other providers including nurses and social workers, is an important measure for success. ³ , ¹¹ , ¹² , ¹³ , ³⁵

LIMITATIONS

There are several limitations in this study that may influence generalizability of its findings to all PEM physicians. First, there may be selection bias introduced by potential unaccounted confounding factors, as 49% of contacted PEM directors and 28% of contacted PEM fellows participated. While this study had a relatively small sample size (n = 153), it was larger and more diverse than prior studies evaluating similar associations on PEM fellow perceptions and practice of social care. ³⁶ Third, the analyses were unadjusted and may have uncontrolled confounding that cannot be accounted for in this study because of sample size and unmeasured factors such as support staff and funding for social care programs. Finally, there may have been misclassification of access to systematic workflow for social care if participants in the sample population misinterpreted the meaning of this term, which was loosely defined. However, such misclassification may be nondifferential and may drive the results closer to the null. Even with such limitations, the results signal a positive association of organizational factors on provider perspectives and practices of social care.

CONCLUSIONS

This study suggests a greater capacity among pediatric emergency medicine fellows and program directors to deliver social care when social care itself is integrated into ED organizational operations. Access to a social care systematic workflow to consistently provide social risk awareness in the ED appears to be an important factor associated with pediatric emergency medicine physician perspective and practices of social care. Best practices for ED‐based social risk awareness and assistance practices are emerging along with models for integrating social care into ED workflow. While adoption across U.S. pediatric EDs has been modest, an understanding of physician perception and practices of social care may help drive greater development of organizational operations around social care. We continue our investigation on ED‐based social care with an analysis of training factors and physician social care practice as well as organizing a national task force charged with development of a consensus‐backed social care training toolkit for health care organizations.

AUTHOR CONTRIBUTIONS

Raymen Rammy Assaf conceptualized and designed the study, developed and piloted the survey, led acquisition of data, and drafted and revised the manuscript. Ryan David Assaf led data curation, executed the statistical analyses and interpretation of data, and drafted and revised the manuscript. Hannah Barber Doucet designed the study, developed and piloted the survey, and revised the manuscript. Danielle Graff contributed to the design and implementation of the research, was involved in planning and supervised the work, and assisted in the overall analysis/interpretation and to the critical revisions of the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

CONFLICT OF INTEREST

The authors have no example conflicts of interest to disclose.

Assaf RR, Assaf RD, Barber Doucet H, Graff D. Pediatric emergency department organization and social care practices among U.S. fellowship programs. AEM Educ Train. 2022;6:1‐8. doi: 10.1002/aet2.10791